How to deal with extremely high spam levels

[Greg McCann]

On 6/22/2004 at 1:32 PM Bob Vincent <bogofilter at bobvincent.org> wrote:

>However, at this rate, it will take nearly a year to collect enough
>non-spam to run bogotune.

Bob,

Here are a couple of tricks that have helped me to keep ham and spam scores automatically updated without spending time correcting the mistakes I used to get from using the -u option.  In my .procmailrc file I have a list of a dozen or so trusted headers that automatically get scored as ham.  For example...

# to keep ham current, autoscore some trusted mail as ham
:0c:bogofilter.lck
* 1^0 ^List-Id: <goodlist at goodlists.com>$
* 1^0 ^To: goodcustomer at yourfavoritestore.com$
* 1^0 ^From: <notaspammer at niceguys.com>
| bogofilter -n -d/home/bogofilter

Not an instant fix for your problem, but it can save a bit of work in the long run.  It also cuts down on false positives if this automatic ham registration comes before bogofilter scoring in .procmailrc.

In addition to this, there are over 50 spamtrap addresses on my mail server.  These addresses receive 1500 spams a day, and email to any of them automatically gets directed to bogofilter for spam registration before being discarded.  Since chances are good that a spamtrap will receive a given spam before I do, bogofilter reacts quickly to new spam campaigns and the endless spelling variations of the V-word, protecting me from new spam without me ever seeing it.  

By the way, none of these spamtrap addresses were created on purpose - I found from my mail logs that certain invalid addresses at my domain were regularly receiving lots of spam, so I took advantage of this and turned them into spamtraps.

HTH,


Greg McCann